Báo cáo khoa học: "Guiding Semi-Supervision with Constraint-Driven Learning" potx

... time periods and scored articles with log-likelihood ra- tio scores. Kanhabua and Norvag (2008; 2009) ex- tended this approach with the same model, but ex- panded its unigrams with POS tags, collocations, and ... version with the same features. We used a MaxEnt model and evaluated with the same filtering methods based on POS tags and tf-idf scores. The model performed best on the dev...

Ngày tải lên: 07/03/2014, 18:20

... with a naive Bayes classifier, the sense s assigned to an example with features f 1 , . . . , f n is chosen so as to maximize: p(s) n  j=1 p(f j |s) In our domain adaptation study, we start with ... of the synsets are simply labeled with “factotum”, indicating that the synset does not belong to a specific domain. 2 25 nouns have at least one synset labeled with the listed domain labe...

Ngày tải lên: 08/03/2014, 02:21

... in- dicating being a name or not. Without these con- straints, t-table entries for names with low frequency tend to be flat and word alignments can be chosen randomly without sufficient statistics or ... assume that a name is produced by a name with a high probability but by a non-name with a low probability, i.e. P(F = E) >> P(F = E), proper names with low counts then are encour...

Ngày tải lên: 20/02/2014, 12:20

... falls with the supertag complexity, the correlation is not absolute: It seems markedly easier to predict su- pertags with complements but no direction infor- mation (C) than supertags with direction ... at all. We combine the label with the attachment direction to obtain composite supertags. The sequence of supertags describing the example sentence would then begin with EXPL/R S/N A...

Ngày tải lên: 23/03/2014, 18:20

... Czech Republic, June 2007. c 2007 Association for Computational Linguistics Guiding Semi-Supervision with Constraint-Driven Learning Ming-Wei Chang Lev Ratinov Dan Roth Department of Computer ... a way to label, along with the current learned model, unlabeled examples. Given a small amount of la- beled data and a large unlabeled pool, our frame- work initializes the model with the...

Ngày tải lên: 31/03/2014, 01:20

... PR with the V32I mutation; PR I50V , PR with the I50V mutation; PR I54M , PR with the I54M mutation; PR I54V , PR with the I54V mutation; PR I84V , PR with the I84V mutation; PR L90M , PR with ... compared with PR WT –SQV. (B) PR I50V –APV compared with PR I50V –SQV. (C) PR I54M –APV compared with PR I54M –SQV. (D) PR I54V –APV compared with PR I54V –SQV. (E) PR I84V –APV...

Ngày tải lên: 18/02/2014, 04:20

... R3. The PCR product with a size of  1.6 kb was digested with BamHI to obtain a fragment between the two copper-binding site coding regions and ligated to pBlueScript pKSII(+) with T4 DNA ligase ... R3-1501 – than R3- 0337 – . This pattern agrees with the presence of two different enzymes with overlapping activities. Purification of two enzymes with different affinities for monophen...

Ngày tải lên: 19/02/2014, 07:20

... a pattern with which to expand it and then filling out that pattern with words drawn from a unigram distribution. This kind of coupled generative process can be naturally formalized with tree ... gen- erate MR and NL associated with MR rules m -v (top). Transducer rule 2 selects MR rule r from the MR gram- mar. Rule 3 simultaneously writes the MR associated with rule m and chooses an...

Ngày tải lên: 19/02/2014, 19:20

... the SRH. We show the performance of GIZA++ and HM- BiTAM with the SRH in the lines entitled with SRH” in Table 1. The GIZA++ and HM-BiTAM with the SRH slightly outperformed the standard GIZA++ ... alignment model. We formulate a syn- onym pair generative model with a topic variable and use this model as a regularization term with a bilingual word alignment model. The topic vari- able...

Ngày tải lên: 20/02/2014, 04:20

... infrastructure to sup- port the creation of, experimentation with, and evaluation of coreference resolvers. Reconcile was designed with the following seven desiderata in mind: • implement the ... present). The basic architecture of the system includes five major steps. Starting with a corpus of docu- ments together with a manually annotated corefer- ence resolution answer key 1 , Reconc...

Ngày tải lên: 20/02/2014, 04:20